4. Discussion
Besides the strongly limited amount of available research data on the susceptibility of mastitis pathogens obtained from German dairy farms, the comparison is further complicated by various methods available for susceptibility testing. The performance of susceptibility testing in accordance with internationally approved guidelines as published by the German Institute for Standardization (referred to as DIN (Deutsches Institut für Normung e. V.)) ensures the high reproducibility of the test results. While the agar disk diffusion test represents a qualitative method, not only being inexpensive to implement and easy to use in practice, the broth microdilution method is a quantitative method considered as the “gold standard” for susceptibility testing due to its complexity and accuracy [
18]. Results as well as derived interpretations differ with regard to the method used for susceptibility testing, this making direct comparison of these two methods quite difficult [
19,
20].
Within the present study, investigated coliforms did not differ in terms of inhibition by cefquinome and amoxicillin/clavulanic acid, whereas marbofloxacin and sulfamethoxazole/trimethoprim MIC
90 of
E. coli were at least two dilutions higher than those determined for
Klebsiella species. More striking differences were apparent concerning cefoperazone MIC
90 values of NAS (8 µg/mL) and
S. agalactiae, as well as
S. dysgalactiae (0.25 µg/mL, respectively), differing by five dilution levels. While
S. dysgalactiae might be considered as susceptible to most tested antimicrobials due to the comparatively low MIC values,
S. uberis frequently exhibited the highest MICs among examined Gram-positive pathogens. Even though oxacillin and cloxacillin are both semisynthetic β-lactams belonging to the group of isoxazolyl penicillins [
21], most Gram-positive pathogens were inhibited at lower oxacillin concentrations.
Previous German studies completely inhibited
S. agalactiae at a penicillin concentration of ≤0.06 µg/mL [
22,
23] and cefquinome concentrations of either ≤0.06 µg/mL [
23] or ≤0.125 µg/mL [
24]. Amoxicillin/clavulanic acid MIC
50/90 corresponded to the same concentration in trials of Minst et al. (2012) (≤0.03 µg/mL) [
22] and the national resistance monitoring program (0.125 µg/mL) [
23]. Wente and colleagues (2016) investigated lower cephapirin MIC
50/90 values (≤0.125 µg/mL) [
24], whereas our results of cefoperazon and oxacillin MIC
90 values are consistent with the latest resistance monitoring [
23]. Therefore, previous studies obviously differed from our results concerning MIC values of cephapirin (MIC
50/90: 0.25/0.5 µg/mL) and amoxicillin/clavulanic acid (MIC
50/90: 0.5 µg/mL), whereas results of remaining antimicrobials were largely consistent.
Our findings concerning
S. dysgalactiae isolates are in agreement with former trials regarding cefquinome and penicillin, since at least 90% were inhibited at ≤0.015 µg/mL within various studies [
22,
23,
25]. In contrast, Wente et al. (2016) examined a higher cefquinome MIC
90 of 0.25 µg/mL, whereas cephapirin MIC values were in line with our study [
24]. Oxacillin MIC
90 determined within a previous study was 0.06 µg/mL [
23] and thus one dilution lower than our study, whereas results of Tenhagen et al. (2006) were in agreement with our survey [
25]. Amoxicillin/clavulanic acid MIC
50 and MIC
90 determined by Tenhagen et al. (2006) was 0.125 µg/mL, respectively [
25], while other German findings (MIC
50/90: ≤0.03 µg/mL) differed by at least two dilutions from our results [
22,
23].
MIC values of penicillin and oxacillin determined for
S. uberis in the present study are in line with those in a former study [
23]. Minst et al. (2012) investigated penicillin MIC
50 and MIC
90 of
S. uberis at concentrations of 0.03 µg/mL and 0.125 µg/mL, respectively [
22]. German studies varied widely regarding cefquinome MIC
90 values of
S. uberis that were determined as 0.125 µg/mL [
25], 0.5 µg/mL [
24] and 0.25 µg/mL [
23], respectively. Similar differences were also apparent for amoxicillin/clavulanic acid MIC
90 values: 0.25 µg/mL [
25], 0.125 µg/mL [
22], 0.5 µg/mL [
23], respectively.
The MIC
50/90 of
S. aureus agree regarding cefquinome [
25,
26], cefoperazone [
26] and amoxicillin/clavulanic acid [
25], whereas the MIC
90 values of our trial were one dilution higher concerning oxacillin [
25,
26] and amoxicillin/clavulanic acid [
26]. Wente and colleagues (2016) examined higher MICs of
S. aureus regarding cefquinome (MIC
50/90: 1/2 µg/mL) and cephapirin (MIC
50/90: 0.25/1 µg/mL) [
24]. Penicillin MIC values investigated within the latest resistance monitoring reveal distinct differences compared to our study. Even though MIC
50 (0.03 µg/mL) predominantly agree with our results (≤0.06 µg/mL), MIC
90 was determined as 16 µg/mL [
26].
In comparison with the latest monitoring, results of NAS were in agreement with our study regarding MIC
50/90 values of cefquinome and amoxicillin/clavulanic acid [
26]. Wente et al. (2016) determined higher cefquinome MIC
50 and MIC
90 values of 1 µg/mL and 2 µg/mL, respectively [
24]. Additionally, cephapirin MIC
90 (1 µg/mL) was one dilution higher [
24], whereas oxacillin and penicillin MIC
90 [
26] were in line with our research study. Results for cefoperazone also differed, since an MIC
50 and an MIC
90 of 1 and 4 µg/mL, respectively, were previously determined [
26].
Susceptibility of
Klebsiella species remained at a constant level within the latest resistance monitoring programs. MIC
90 of our study differed by one to four dilutions, since the BVL (Bundesamt für Verbraucherschutz und Lebensmittelsicherheit) determined MIC
90 of 0.06 µg/mL (cefquinome), 4 µg/mL (amoxicillin/clavulanic acid), 0.06 µg/mL (marbofloxacin) and 0.25 µg/mL (sulfamethoxazole/trimethoprim) [
23,
26].
Whereas cefquinome MIC
90 of
E. coli are in line with a former study [
23], Wente and colleagues (2016) examined cefquinome MIC
90 of 32 µg/mL [
24]. Moreover, cefquinome MIC
90 determined within the resistance monitoring also varied over the years: 0.125 µg/mL (2010), 8 µg/mL (2012), 0.125 µg/mL (2014) and 0.5 µg/mL (2016) [
23,
26,
27,
28]. Interestingly, similar strong fluctuations were apparent concerning marbofloxacin (0.03 µg/mL (2012), 0.5 µg/mL (2014), 0.06 µg/mL (2016)) and sulfamethoxazole/trimethoprim (0.25 µg/mL (2010, 2012), ≥64 µg/mL (2014, 2015)) [
23,
26,
27,
28].
The resistance monitoring achieved results from susceptibility examinations from various laboratories located in different German federal states. Minst et al. (2012) already detected higher resistance rates in districts with a high density of dairy farms and assumed that this might be due to a locally increased antimicrobial use [
22]. The current study focused solely on mastitis pathogens obtained from dairy farms in Northern Germany, which is characterized as a particularly densely populated region of livestock [
10,
29]. Data on dispensed antimicrobials in veterinary medicine indicated a considerably higher amount in the North of Germany, especially in the federal state of Lower Saxony [
6]. Although resistance of staphylococci to antimicrobials was frequently higher in regions with an above-average intensity of dairy cattle in Northern Germany, a previous research study concluded that the density of swine or cattle populations is not associated with the frequency of resistant bacteria in a region [
29]. In contrast, Tenhagen et al. (2018) concluded that prevalence of MRSA seems to be related to the density of livestock and might thus be elevated in densely populated regions. Furthermore, Tenhagen and colleagues also found a positive correlation between prevalence of MRSA in bulk tank milk and the respective herd size. In this regard, 13.3% of milk samples obtained from herds with >80 cows were tested positive for MRSA, whereas the amount was 7.3% in herds having <80 cows [
10]. The current study included mastitis pathogens from dairy herds with a herd size ranging from 55 to 2500 dairy cows. To investigate differences in susceptibilities in relation to herd size, the participating farms were divided into 30 small dairy farms (<92 cows) and 28 large ones (>92 cows). The limiting number of 92 cows was used as this number represents the current average herd size in the Northern German region [
30].
Table 8 and
Table 9 display the MIC values for Gram-positive and Gram-negative pathogens by differentiating between small and large dairy herds, respectively.
S. agalactiae was isolated from small dairy farms only, so the results of this mastitis pathogen are not presented.
The differences for both Gram-positive as well as Gram-negative pathogens almost exclusively refer to investigated MIC90 values. A higher MIC50 value of 1 µg/mL was detected for isolates from larger farms concerning the following bacteria-antimicrobial-combinations: Klebsiella species (sulfamethoxazole/trimethoprim), S. aureus (cefquinome) and NAS (oxacillin). Regarding the tested NAS isolates, no additional differences were apparent between small and large farms, whereas MIC90 values of S. aureus isolates from large dairy farms were one concentration level lower (penicillin: 0.25 µg/mL; amoxicillin/clavulanic acid: 0.5 µg/mL) or higher (cephapirin: 0.5 µg/mL; cefalexin/kanamycin: 4 µg/mL) than the respective values determined for the isolates from small dairy farms. In contrast, the investigated Streptococci from small farms were frequently inhibited at elevated antimicrobial MIC90 values. In the case of the tested S. dysgalactiae isolates, this was obvious for cefoperazone, cloxacillin, oxacillin and cefalexin/kanamycin; whereas S. uberis isolates from herds >92 cows showed higher MIC90 values for all antimicrobial agents included in the study. In this respect, cloxacillin MIC90 (16 µg/mL) and cefalexin/kanamycin MIC90 (8 µg/mL) of S. uberis from small dairy farms represented the greatest difference from the isolates obtained from large farms (oxacillin MIC90: 2 µg/mL; cefalexin/kanamycin MIC90: 1 µg/mL), which differ in three concentration levels. There are also noticeable differences in the MIC90 values of the antimicrobials tested against Gram-negative bacteria. Klebsiella species isolated from large farms were therefore inhibited at MIC90 values of cefquinome (0.5 µg/mL), marbofloxacin (0.125 µg/mL) and sulfamethoxazole/trimethoprim (4 µg/mL) that were one concentration level higher than respective MIC90 from Klebsiella isolates from small herds, respectively. The amoxicillin/clavulanic acid MIC90 values of investigated Klebsiella were determined as 1 µg/mL (small farm) and 16 µg/mL (large farm). When discussing the results of the Klebsiella species, the different number of bacterial isolates from small (n = 12) and large herds (n = 40) must be taken into consideration. Due to the substantially smaller quantity of isolates from small farms, a single isolate represented 8.3% of the total and consequently has a higher impact on MIC50/90 values which refer to the total amount of tested bacteria. Regarding E. coli isolates, the MIC90 values of cefquinome (0.5 µg/mL) and amoxicillin/clavulanic (16 µg/mL) were higher on small farms, whereas a comparatively higher MIC90 was detected for isolates tested against marbofloxacin (1 µg/mL) and sulfamethoxazole/trimethoprim (32 µg/mL). The difference of five sulfamethoxazole/trimethoprim concentration levels was the highest between E. coli from small (1 µg/mL) and large (32 µg/mL) farms. This was due to four (12.1%) E. coli isolates gained from large farms, that were not inhibited within the tested concentration range and thus had MICs of >16 µg/mL. The remaining 29 isolates (87.9%) obtained from large farms as well had MICs of concentrations ranging from 0.25 µg/mL to 2 µg/mL for sulfamethoxazole/trimethoprim. Taking everything into consideration, a general statement about the occurrence of less susceptible mastitis pathogens existing on larger dairy farms cannot be confirmed on the basis of this study. Even though antimicrobial MICs were higher in isolates from larger farms, especially in coliforms, the MIC90 of streptococci were exclusively higher in large dairy herds, while NAS showed no differences, except for oxacillin MIC50. However, more investigations including a greater bacterial number are needed before an accurate statement can be made.
The susceptibility of Gram-positive mastitis pathogens obtained from dairy farms located in Northern Germany was previously described as favorable, whereas Gram-negative isolates exhibited varying degrees of resistance regarding all tested antimicrobials. Cephalosporins were considered to be effective against staphylococci and streptococci mastitis. However, penicillin represented the drug of choice for treating clinical mastitis induced by streptococci [
31]. By establishing special guidelines, various countries aim to reduce the use of antimicrobials in order to counteract the development of resistance. This especially refers to those antimicrobial agents that were declared as “critically important” due to their major relevance in human medicine [
32]. In Germany, obligatory guidelines have recently been introduced for the use of certain antimicrobial substances. The microbiological analysis of milk samples as well as susceptibility testing of present pathogens are thus required whenever use of third- and fourth-generation cephalosporins or fluorochinolones is considered for antimicrobial treatment [
33]. In comparison to Germany, several Nordic countries such as Sweden, Norway, Denmark and Finland already created a common strategy for treating clinical mastitis several years ago. According to the jointly created “Nordic Guidelines for Mastitis Therapy” a general restrictive use of penicillin is pursued by an overall reduction in cephalosporins and quinolones as much as possible [
12]. Chehabi et al. (2019) support the use of penicillin as the drug of first choice on Danish dairy farms based on its efficacy in in vitro susceptibility examinations of mastitis pathogens. Therefore, streptococci were classified as highly susceptible to penicillin, whereby results of
S. uberis (MIC
50/90: ≤0.06/0.25 µg/mL),
S. dysgalactiae and
S. agalactiae (MIC
50/90: ≤0.06/≤0.06 µg/mL) were largely consistent with our results. Whereas penicillin MIC values of Danish NAS were ≤0.06 µg/mL (MIC
50) and 0.5 µg/mL (MIC
90), results of
S. aureus were higher for penicillin MIC
90 (2 µg/mL), while penicillin MIC
50 was ≤0.06 µg/mL [
34]. A lower penicillin MIC
90 of Danish
S. aureus was previously reported at concentrations of 0.25 µg/mL, this being identical with results of
S. aureus isolates from Iceland and Switzerland (MIC
50/90: ≤0.06/0.25 µg/mL). Penicillin MIC values of
S. aureus from England, Finland, Ireland, Sweden and the United States were in accordance with our results detected at concentrations of ≤0.06 (MIC
50) and 0.5 µg/mL (MIC
90) [
35]. Oxacillin MIC
50/90 of
S. aureus investigated in our study was determined at concentrations of 0.25 µg/mL and 1 µg/mL, respectively. In comparison to results of previous surveys from England, Finland, Iceland, Ireland, Norway, Sweden, Switzerland and the United States, the oxacillin MIC
90 of
S. aureus was also 1 µg/mL, while MIC
50 was 0.5 µg/mL.
S. aureus isolates from Denmark were inhibited at lower oxacillin concentrations (MIC
50/90: 0.25/0.5 µg/mL) [
35].
An interesting point is the difference in the MIC values of the isoxazolyl penicillins oxacillin and cloxacillin, which was detected for the majority of mastitis pathogens. Besides the problem of different methods used for susceptibility testing, comparison is also hampered by the use of different antimicrobial substances approved for mastitis treatment in various countries. The differences in the MIC values of oxacillin and cloxacillin indicate that a comparison of different antimicrobial agents belonging to the same antimicrobial group is inappropriate, such as comparing the third-generation cephalosporins cefoperazone (approved in Germany) and ceftiofur (approved in the United States). Most of the research studies only included oxacillin for detecting the in vitro efficacy of isoxazolyl penicillins against bacteria. An international study by de Jong et al. (2018) investigated the susceptibility of
S. aureus and NAS against both oxacillin and oxacillin. Whereas NAS exhibited MIC
50 values of cloxacillin and oxacillin at concentrations of 0.5 µg/mL, the MIC
90 differed for cloxacillin (2 µg/mL) and oxacillin (1 µg/mL). In contrast,
S. aureus was inhibited at cloxacillin MIC
50 and MIC
90 values of 0.25 µg/mL and 0.5 µg/mL, respectively, while MICs of oxacillin were one concentration level higher (MIC
50/90: 0.5/1 µg/mL) [
36]. Whereas the findings of NAS support the statement of higher cloxacillin MICs, the results of
S. aureus did not support this hypothesis. De Jong et al. (2018) summarized the results of susceptibility testing from nine European countries: Belgium, the Czech Republic, Denmark, France, Germany, Italy, the Netherlands, Spain and the United Kingdom. Results of de Jong et al. (2018) were different to those of our study for
S. aureus (amoxicillin/clavulanic MIC
50: 0.125 µg/mL; cefquinome MIC
90: 0.5 µg/mL) as well as for NAS (cephapirin MIC
50/90: 0.125/0.25 µg/mL), whereby the MICs published by de Jong et al. (2018) were one concentration level lower. The remaining MIC
50/90 of amoxicillin/clavulanic acid, cephapirin and cefquinome did not differ between our results and those of de Jong et al. (2018) [
36]. Concerning
S. dysgalactiae differences to our study were detected in MIC
90 values only, as MIC
90 values of ≤0.03 µg/mL (amoxicillin/clavulanic acid), 0.06 µg/mL (cephapirin) and 0.015 µg/mL (cefquinome) were previously described [
36]. Compared to our result, the MICs of
S. uberis were lower in the case of amoxicillin/clavulanic acid (MIC
50/90: 0.25/0.5 µg/mL), cephapirin (MIC
90: 0.5 µg/mL) and cefquinome (MIC
50/90: 0.125/0.25 µg/mL) [
36]. Both
Klebsiella species and
E. coli previously exhibited cefquinome MIC
50/90 of 0.06/0.125 µg/mL and thus differed by one (MIC
50) and two (MIC
90) concentration levels from our trial [
36]. Amoxicillin/clavulanic MIC
90 published by de Jong and coworkers (2018) was 8 µg/mL regarding both
Klebsiella species and
E. coli. Moreover, 90% of
E. coli isolates were inhibited at a marbofloxacin concentration of 0.06 µg/mL (MIC
90), representing a three concentration level lower value compared to our study. Marbofloxacin MIC
50/90 against
Klebsiella species were formerly detected as 0.06 µg/mL and differed concerning both values from our investigations (MIC
50/90: 0.03/0.125 µg/mL) [
36]. As a conclusion of the comparison with previous German and international studies, it can be stated that staphylococci predominantly show the same susceptibility patterns to most antimicrobials. The values for
S. dysgalactiae are mainly limited to low concentration levels and the
S. uberis isolates in our study also deviates strongly from the former German and international results. The high MIC values compared to cefquinome could possibly be due to the increased use of these antimicrobials on German dairy farms. However, this is not supported in the case of other pathogens which did not show increased cefquinome MICs and the fact that MIC values of other antimicrobials (e.g., cephapirin) tested against
S. uberis (
Table 3) were also elevated, although they are not very relevant in mastitis therapy in Germany.
The favorable results of the Danish, Norwegian and Swedish mastitis pathogens could be due to a highly restrictive use of antimicrobials in these countries. Prudent use—in order to minimize the selective pressure—should therefore be a mandatory requirement whenever the use of antimicrobial agents is considered. In the context of antimicrobial reduction in mastitis treatment, the possibility of alternative treatment methods should be considered. An alternative treatment method for mastitis cases could be the use of bacteriophages as therapeutic agents (bacteriophage therapy). Bacteriophages (also called phages) are viruses that only target prokaryotic cells, mainly of one bacterial species, for inserting their DNA or RNA for propagation. After propagation, phage-coded enzymes induce the lysis of the bacterial cell (lytic propagation cycle) which leads to the release of next generation phages that are able to infect new host cells [
37]. In a mouse model conducted by Capparelli et al. (2007), a bacteriophage therapy was able to achieve a complete reduction in the
S. aureus in mice that were simultaneously infected with bacteria and phages [
38]. In another study,
S. aureus was previously isolated from dairy cows suffering from mastitis and used to induce mastitis in mice as well. Phage therapy was able to reduce the bacterial counts as well as the clinical degree of the disease. The authors considered the use of phages for an alternative therapy option for the treatment of bovine mastitis, but also mentioned the non-comparability of mastitis in mice and cows due to anatomical and physiological conditions [
39]. The fact that lytic results investigated within a mouse model cannot directly be transferred to a dairy cow suffering from mastitis has already been demonstrated by Gill et al. (2006) [
40]. In this former study, solely 16.7% (3/18) (
p > 0.05) of the quarters infected with
S. aureus achieved a bacteriological cure after a five-day treatment with an infusion of a high concentration of bacteriophage K. The authors assumed that an effective concentration of phages in the mammary gland of cattle could not be achieved in raw milk, among other things, due to an inhibition of the bacteriophages to bind to the host cell surface [
40]. The investigations by Gill et al. (2006) refer to cows suffering from subclinical mastitis and, to the best of our knowledge, clinical trials are still lacking.
Moreover, a therapy option is the application of non-antimicrobial agents like products containing proteolytic enzymes such as chymotrypsin and trypsin [
41] or homeopathic remedies [
42]. The former was tested in cows suffering from mild to moderate clinical mastitis, while the latter was administered to chronically infected animals. In both studies, the authors concluded that the treatment could be considered as a possible alternative to antimicrobial therapy. Nevertheless, the efficacy of both products for treating severe clinical mastitis cases may be considered doubtful [
41,
42].
Another possibility for avoiding the extensive use of antimicrobials could be the application of vaccines. In several trials, the application of vaccines was able to cause a reduced severity of the clinical signs accompanied with mastitis cases caused by
S. uberis and
E. coli [
43,
44]. Furthermore,
S. uberis-specific vaccines were beneficial concerning the reduction in somatic cell count and bacterial count [
44]. While Schukken et al. (2014) achieved a reduction in both the incidence and prevalence of intramammary infections with staphylococci [
45], previous observations by Tenhagen et al. (2001) were contrary. Therefore, the application of a herd-specific vaccination in order to prevent
S. aureus-induced mastitis in heifers was not successful, since neither the prevalence of intramammary infections nor the incidence of clinical mastitis was significantly approved [
46]. The broad range of alternative treatment options (bacteriophages, vaccination, enzymes, homeopathic agents) must be, as already recommended by the majority of authors, further researched in order to establish these methods as effective alternatives in mastitis therapy. Nevertheless, these points might be beneficial in the pursued aim of reducing the antimicrobial use on dairy farms in the future.
The aim of the present study was rather to focus on mastitis pathogens having comparatively higher MIC values, so that success of an antimicrobial therapy might be predicted as doubtful. While in the case of most mastitis pathogens, this only referred to individual isolates, an evident proportion of
E. coli isolates were not inhibited within tested concentration ranges (
Table 7). This leads to the assumption that individual resistant isolates are present on German dairy farms. However, the final classification of a pathogen as resistant against a certain antimicrobial agent is performed by using clinical breakpoints. Regarding the performance of susceptibility testing, it is obligatory that the used methodology and interpretive criteria are performed following the same guidelines [
20]. The Clinical and Laboratory Standards Institute (CLSI) established clinical breakpoints for only three antimicrobial agents referring to the indication of bovine mastitis: ceftiofur, penicillin/novobiocin and pirlimycin [
47]. While ceftiofur and penicillin/novobiocin are even not approved for the intramammary application in Germany, pirlimycin is not relevant for mastitis treatment according to current market shares (unpublished data provided by the German Consumer Research Company (GfK)). Due to the lack of specific clinical breakpoints, identifying resistant pathogens is frequently based on breakpoints established for a different animal species, different indication or even established for human medicine. A transmission of these breakpoints is not appropriate and can cause misinterpretations of the results [
20,
48]. Moreover, in some trials, a classification of pathogens as susceptible and resistant did not result in the enhanced success of a therapy that was derived from the test results [
49,
50,
51]. While the β-lactam antimicrobials are excreted to a high degree in the urine and thus reach elevated concentrations there, the clinical breakpoints related to urinary tract diseases were generally set higher for these antimicrobials [
48]. Undoubtedly, the conditions of the urinary tract are not directly transferable to the mammary gland of a dairy cow and presumptive pathogens present in the bovine udder. In order to avoid the pitfalls associated with the use of improper clinical breakpoints, we focused rather on distribution patterns and MIC values. Changes in MIC values can therefore be detected by a comparison with former susceptibility trials, especially by considering regional differences. An assessment and discussion based on MIC values might thus be an appropriate alternative to strictly classifying pathogens as susceptible or resistant, supported by the fact that the statement is even questionable. However, the establishment of mastitis-specific clinical breakpoints is an indispensable tool for accurately assessing resistances and therapy decisions in future.